Investigators of the Piedmont Health Survey of the Elderly (Established Populations for Epidemiologic Studies of the Elderly/Duke) propose to perform a ten-year, fourth in-person follow-up of the North Carolina Piedmont cohort initially interviewed in 1986/87. Of the original 4,162 community-dwelling elders 65+ years of age, equally distributed between urban versus rural residence and Black versus non-Black race/ethnicity, it is anticipated that 2,060 subjects or proxy respondents will be available for interview. The purpose of this follow-up study is to obtain information on four primary outcome variables (cognitive status, depression, functional status, and mortality) and four primary independent variables (social support, social class, social location, and chronic illness). By using data from four in-person interviews across the ten years of follow-up, investigators will be able to characterize trajectories of the major dependent variables and their relationship to the independent variables over time. This will provide valuable information about change and stability, as well as heterogeneity in the dynamics of change. The major goal is to examine the relationships between social factors and chronic disease on the one hand, and health outcomes on the other. In order to examine both intraindividual change and the effects of variables that differ across individuals, hierarchical linear models (HLM) will be the primary method of analysis. Data gathered at four points in time across ten years will permit sophisticated analytic techniques to be applied to a data set of superior quality, from a diverse racial/ethnic and geographically- distributed sample which exhibits appreciable variation over time in the areas listed above.